filling empty variables conditional on non empty variables

[ilikesas]

Hi,

suppose i have the following table:

obs	pct shares	pct stock
1	30	70
2	40
3		55
4

The sum of the 2 pct shares and pct stock should equal to 100 for every observation. In the cases when one of the two is missing, I need to find the other one by subtracting the non missing value from 100. In the cases when both are present or missing leave as is.

So the desired table is the following:

obs	pct shares	pct stock
1	30	70
2	40	60
3	45	55
4

Thank you!

[stat_sas · Accepted solution]

data want;

set have;

if pctshares=. then pctshares=100-pctstock;

if pctstock=. then pctstock=100-pctshares;

run;

[Patrick]

Amending the code from stat@sas to only calculate values "when one of the two is missing".

data want;

  set have;

  if cmiss(pctshares,pctstock)=1 then

    do;

      if pctshares=. then pctshares=100-pctstock;

      else pctstock=100-pctshares;

    end;

run;